bankr-signals
Pass
Audited by Gen Agent Trust Hub on Mar 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [COMMAND_EXECUTION]: The skill includes a bash script
scripts/publish-signal.shthat executesnode -efor message signing andcurlfor API communication.HEARTBEAT.mdincludes shell commands and an inline Python snippet for processing leaderboard data.- [DATA_EXFILTRATION]: The skill requires sensitive inputs including aPRIVATE_KEYenvironment variable and an API key stored in~/.clawdbot/skills/bankr/config.json. These are used to authenticate with the vendor's services (api.bankr.botandbankrsignals.com).- [PROMPT_INJECTION]: The agent processes signal data from an external feed to influence trading actions. - Ingestion points: Signal data is fetched from
https://bankrsignals.com/api/feedas described inHEARTBEAT.md. - Boundary markers: The skill lacks explicit boundary markers for the ingested signal reasoning text, though it encourages structural filtering.
- Capability inventory: The skill uses
scripts/publish-signal.shto execute signing and network operations via subprocesses. - Sanitization: The skill suggests validating signals against on-chain transaction hashes (
txHash) and provider performance metrics as a mitigation strategy.
Audit Metadata